Pyknometers (what hydrometers emulate) revisited!

[Marshbrewer]

Perhaps I'm misunderstanding, but I still don't understand how you account for the weight of the yeast in a fermented sample.

[PeeBee]

Perhaps I'm misunderstanding, but I still don't understand how you account for the weight of the yeast in a fermented sample.

Thanks for the brain exercise I probably need it!

I was going to reply with something cocky, like "what do people do with hydrometers" (as the subject title states hydrometers only emulate what a pyknometer does for a past time with no cheap electronics). But that would have only been admitting "I don't know". ***

The reason there is no need to account for all the suspended yeast in a fermented sample is because of the nature of the yeast - it's suspended. Like submarines, they share the same overall density as their surroundings, so there is no need to account for suspended yeast at all.

Yeast must have a slightly higher density else they wouldn't sink at the end of fermentation (CO2 evolution and convection currents must help keep them suspended otherwise) but the difference will be well outside our (homebrewers) ability to measure. The yeast probably undergoes some physical changes to accelerate "sinking" at the end of life or food supply, but that's well beyond my knowledge; anyway, I think we all avoid scraping the bottom of the fermenter for a hydrometer sample so that's not an issue.



*** On this subject: I reckon I'm just part of the vanguard moving against hydrometers for homebrewing. I predict it won't be long before the demise of the hydrometer; they'll probably be replaced with whatever passes as a "mobile phone app" in the near future. If we can avoid attempted annihilation of the human species meantime?

[Marshbrewer]

The reason there is no need to account for all the suspended yeast in a fermented sample is because of the nature of the yeast - it's suspended. Like submarines, they share the same overall density as their surroundings, so there is no need to account for suspended yeast at all.

Yeast must have a slightly higher density else they wouldn't sink at the end of fermentation (CO2 evolution and convection currents must help keep them suspended otherwise) but the difference will be well outside our (homebrewers) ability to measure. The yeast probably undergoes some physical changes to accelerate "sinking" at the end of life or food supply, but that's well beyond my knowledge; anyway, I think we all avoid scraping the bottom of the fermenter for a hydrometer sample so that's not an issue.

I missing something here. If yeast have the same density as wort, then they can't have the same density as wort plus alcohol. Yet they remain suspended. I'm probably overthinking this

[Eric]

That is the point, yeast can and do, just at different times. If they couldn't, it would take a lot longer to make beer because they'd all be on the top at the start and on the bottom quickly if any of them did some work. Like all living cells or beings, they adjust to survive in their environment.

[PeeBee]

I was getting in a bit of a tizz with me Maths. But I needn't have been worrying, I was just not explaining myself very well (that's explaining myself to myself, not you, although I don't think my lack of explanation helped anyone). If I appear to have a condescending tone, that is also for me not you!

I started alright:

... Divide the NETT weight by the "magic number" to get the bottle's VOLUME. e.g. 99.49 / 0.9982 = 99.67 millilitres. ...

... at 20°C the water sample will weigh a smidge lighter (in grams) than the bottle's volume (in millilitres). By dividing by the density of water at 20°C the result is the volume of the bottle (density of sample has been converted to 1.000g/ml), or what it would have weighed if the sample were at 4°C. Note: This only works for water.

But after that it gets a bit abstract! Not surprising really, I start applying "tricks".

At the moment I only work at 20°C. This I'm working at changing (for now the "magic number" is fixed at 0.9982, and the bottle's "Tare" weight altered to compensate). But I fiddle the "Tare" weight to get an "imaginary" bottle volume of 100.00ml exactly. This is where I (and perhaps you) start to lose it!

I've fiddled the bottle's volume, but because of the "trick Tare weight" I'm fiddling the weight of sample by near-enough the same magnitude; I'm pretending the weight of the sample is heavier than it actually is. It was this pretence that was throwing me off-track, but I was forgetting the fiddled bottle volume. As the density is effectively a ratio, the weight divided by the volume, it makes no difference as long as neither get too far out of sync. I reckon the "trick" is good for "specific gravities" 1.000 to 1.070 (within ±0.0005, depending on the quality of your weighing scales).

The error gets increasingly noticeable as the relative density (specific gravity) gets higher. It is this anomaly (along with temperature compensation) I'm currently working on a fix for.




They (which includes Graham Wheeler!) were getting in a bit of a tizz with the subject too (hydrometer temp correction ... a "pinned" thread ). I believe this was down to considering both temperature compensations as one, which makes for ludicrously complicated formulas (that can't possibly hold true). I'll stick to keeping the two temperature compensations separate.

Some hydrometers will show both temperature calibrations (but it's uncommon). Brewing hydrometers (UK) should be marked "20°C/20°C". "Density Hydrometers" (and the current UK Lab standard) should be marked "20°C/4°C". Most Lab hydrometers are still the obsolete "60°F/60°F", but most Labs switched to oscillating U-tube (densitometers) types long ago.

[PeeBee]

Brewed a batch yesterday; ideal opportunity to try the new pyknometer and weighing regime:

20220901_193005_WEB.jpg (68.28 KiB) Viewed 22064 times

With the beer in the fermenter. the resident TiltPRO Hydrometer gets a look in:

Boddies 1901 AK.JPG (20.25 KiB) Viewed 22064 times

Can't argue with that! Well, the TiltPRO hasn't been calibrated for a few brews because the new PRO design holds its calibration so well. But it was originally configured to another pyknometer, so perhaps no surprise? (Remember this is four decimal places, not three).

So, I'll use this new pyknometer "long-hand" to get a second opinion. Using the "real" GROSS weight less the "real" TARE weight to get the sample's "real" NETT weight:

139.64 - 35.72 = 103.92g

Divide by "real" bottle's VOLUME to get the density of the sample at 20°C:

103.92 ÷ 99.67 = 1.0426g/ml

Divide by water's density at 20°C to get specific gravity:

1.0426 ÷ 0.9982 = 1.0444 SG

Hmm ...

[MashBag]

Perhaps I'm misunderstanding, but I still don't understand how you account for the weight of the yeast in a fermented sample.

Surely the suspend-ability of the water is changing too, as the alcohol is produced. Making it thinner, thus harder for the yeast to stay up?

[PeeBee]

I've ordered a 100ml pyknometer bottle from "samnacu" on EBay (China) for £5.44 + £1.00 shipping + taxes (VAT, collected automatically by Paypal). That's cheap! It'll take 3-4 weeks to arrive. I'll be using my snazzy (expensive to replace) bottle, but I like to know I'm suggesting a practical technique. I've also got my eye on a variety of vials and "tungsten putty" (for carp fishing weights) from which to fabricate a range (3?) of tare weights to account for higher SG samples and different temperature compensations (that's what I'm continuing to work on).

[MashBag]

PeeBee that sounds a smidge like I have offended you, and that wasn't the intention.

I am genuinely interested in this.
I have for a while wanted this level of accuracy, but without spending a fortune of electronics.

I have tried wide band hydrometers which are a faff, expensive and fragile.
Wishing you all the best, and Tbh at a fiver, I am happy to have a go to prove your research.

[PeeBee]

PeeBee that sounds a smidge like I have offended you, and that wasn't the intention.

I am genuinely interested in this.
I have for a while wanted this level of accuracy, but without spending a fortune of electronics.

I have tried wide band hydrometers which are a faff, expensive and fragile.
Wishing you all the best, and Tbh at a fiver, I am happy to have a go to prove your research.

Offended? How'd you work that out? Ignored perhaps, because your question was to "JJSH". I did notice you used the word "thinner" which could be misinterpreted: We were already talking "density" so were you throwing in "viscosity" by saying "thinner"?


But I'd just posted some "actual" results and was I was not getting excited by the close matching. That was because I "accidentally-on-purpose" avoided mentioning the temperature which was a degree warmer in the pyknometer bottle. I'm still working on the implications of temperature and easy compensations. It's straight-forward to correct both temperature configurations with a pyknometer whereas they're fixed on a hydrometer (let's face it, who's bothered to reach for temperature correction tables 'cos their hydrometer sample is a degree or two off 20C?).

The other little issue I'm trying to iron out: I'm now gravitating towards SI Water as a source of figures for water density (at 3+ decimal places, there's a lot of discrepancies between various sources of data, and SG is to at least 3 decimal places). But that density information (and my calculations are based very much on density) is immediately followed by a table on thermal expansion. And working out weight and volume based on density is just very slightly different to thermal expansion:

water density at 20°C = 0.9982g/ml

1g water at 20°C = 1.0028ml

I can't make them add up! And SG (even at 3 DP) would be different if using one rather than the other? 1.0028ml x 0.9982g/ml = 1.000995g, or 1 gravity unit: What am I missing? The table states "Data corrected for buoyancy and for the thermal expansion of the container": What are they babbling about?

[EDIT: Hang on. I'm possibly making the dummy mistake of mixing "density" in grams per millilitre with the ratio "SG" which has no units. Still, I'm not seeing how to correct such an error?]

[PeeBee]

... and Tbh at a fiver, I am happy to have a go to prove your research.

Do remember, the fiver doesn't buy the weighing scales (£30-35). But you won't need those crappy short lived ten-quid "drug dealer" scales for weighing out water salts (I use them to weigh hops too).

[PeeBee]

I've been checking Amazon for suitable weighing scales. A few "midi" sized examples (0.01-500g) for £12-18.

But I wouldn't give much credibility to these scales that you can't even "calibrate". They are probably the same "drug-dealer" scales in a bigger box! And descriptions of "increments" of 0.01g means absolutely nothing! The scales must reliably weigh to within 0.01g and not weigh something different each time you weigh it!

[MashBag]

We are good on scales. I have a few some cheap, and some definite weren't and can be calibrated accurately enough to weight a sparrows fart

[Eric]

If the density of water at 20C is 0.9982g/ml, then 1 gm of water a 20C has a volume of 1/0.9982 = 1.0018 ml.

Where did you find 1.0028? could that be a misprint?

[PeeBee]

Hello Eric! It was in that "SI Metric" site I linked earlier:

...The other little issue I'm trying to iron out: I'm now gravitating towards >SI Water< as a source of figures for water density (at 3+ decimal places, there's a lot of discrepancies between various sources of data, and SG is to at least 3 decimal places). But that density information (and my calculations are based very much on density) is immediately followed by a table on thermal expansion. And working out weight and volume based on density is just very slightly different to thermal expansion ...

It isn't an "authoritative" site (as you will see!), just one I found that was fairly consistent with bits on other sites (most sites I looked at had chunks that weren't consistent with any other site).

I can't imagine the "0.0028" bit was a typo because that would mean the entire table column is at fault.

I think it's something kookie about that "Data corrected for buoyancy and for the thermal expansion of the container" statement? I think I'm okay to just ignore it and stick with the "density" data which is at least working.